MotionClone: Training-Free Motion Cloning for Controllable Video Generation
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866917811680444416 |
|---|---|
| author | Ling, Pengyang Bu, Jiazi Zhang, Pan Dong, Xiaoyi Zang, Yuhang Wu, Tong Chen, Huaian Wang, Jiaqi Jin, Yi |
| author_facet | Ling, Pengyang Bu, Jiazi Zhang, Pan Dong, Xiaoyi Zang, Yuhang Wu, Tong Chen, Huaian Wang, Jiaqi Jin, Yi |
| contents | Motion-based controllable video generation offers the potential for creating captivating visual content. Existing methods typically necessitate model training to encode particular motion cues or incorporate fine-tuning to inject certain motion patterns, resulting in limited flexibility and generalization. In this work, we propose MotionClone, a training-free framework that enables motion cloning from reference videos to versatile motion-controlled video generation, including text-to-video and image-to-video. Based on the observation that the dominant components in temporal-attention maps drive motion synthesis, while the rest mainly capture noisy or very subtle motions, MotionClone utilizes sparse temporal attention weights as motion representations for motion guidance, facilitating diverse motion transfer across varying scenarios. Meanwhile, MotionClone allows for the direct extraction of motion representation through a single denoising step, bypassing the cumbersome inversion processes and thus promoting both efficiency and flexibility. Extensive experiments demonstrate that MotionClone exhibits proficiency in both global camera motion and local object motion, with notable superiority in terms of motion fidelity, textual alignment, and temporal consistency. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_05338 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | MotionClone: Training-Free Motion Cloning for Controllable Video Generation Ling, Pengyang Bu, Jiazi Zhang, Pan Dong, Xiaoyi Zang, Yuhang Wu, Tong Chen, Huaian Wang, Jiaqi Jin, Yi Computer Vision and Pattern Recognition Motion-based controllable video generation offers the potential for creating captivating visual content. Existing methods typically necessitate model training to encode particular motion cues or incorporate fine-tuning to inject certain motion patterns, resulting in limited flexibility and generalization. In this work, we propose MotionClone, a training-free framework that enables motion cloning from reference videos to versatile motion-controlled video generation, including text-to-video and image-to-video. Based on the observation that the dominant components in temporal-attention maps drive motion synthesis, while the rest mainly capture noisy or very subtle motions, MotionClone utilizes sparse temporal attention weights as motion representations for motion guidance, facilitating diverse motion transfer across varying scenarios. Meanwhile, MotionClone allows for the direct extraction of motion representation through a single denoising step, bypassing the cumbersome inversion processes and thus promoting both efficiency and flexibility. Extensive experiments demonstrate that MotionClone exhibits proficiency in both global camera motion and local object motion, with notable superiority in terms of motion fidelity, textual alignment, and temporal consistency. |
| title | MotionClone: Training-Free Motion Cloning for Controllable Video Generation |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2406.05338 |